Dataset Information

Using data-driven rules to predict mortality in severe community acquired pneumonia.

ABSTRACT: Prediction of patient-centered outcomes in hospitals is useful for performance benchmarking, resource allocation, and guidance regarding active treatment and withdrawal of care. Yet, their use by clinicians is limited by the complexity of available tools and amount of data required. We propose to use Disjunctive Normal Forms as a novel approach to predict hospital and 90-day mortality from instance-based patient data, comprising demographic, genetic, and physiologic information in a large cohort of patients admitted with severe community acquired pneumonia. We develop two algorithms to efficiently learn Disjunctive Normal Forms, which yield easy-to-interpret rules that explicitly map data to the outcome of interest. Disjunctive Normal Forms achieve higher prediction performance quality compared to a set of state-of-the-art machine learning models, and unveils insights unavailable with standard methods. Disjunctive Normal Forms constitute an intuitive set of prediction rules that could be easily implemented to predict outcomes and guide criteria-based clinical decision making and clinical trial execution, and thus of greater practical usefulness than currently available prediction tools. The Java implementation of the tool JavaDNF will be publicly available.

SUBMITTER: Wu C

PROVIDER: S-EPMC3974677 | biostudies-literature | 2014

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Using data-driven rules to predict mortality in severe community acquired pneumonia.

Wu Chuang C Rosenfeld Roni R Clermont Gilles G

PloS one 20140403 4

Prediction of patient-centered outcomes in hospitals is useful for performance benchmarking, resource allocation, and guidance regarding active treatment and withdrawal of care. Yet, their use by clinicians is limited by the complexity of available tools and amount of data required. We propose to use Disjunctive Normal Forms as a novel approach to predict hospital and 90-day mortality from instance-based patient data, comprising demographic, genetic, and physiologic information in a large cohort ...[more]

PMID: 24699007

Similar Datasets

Project description:IntroductionCommunity-acquired pneumonia (CAP) requiring intensive care unit (ICU) treatment commonly causes acute respiratory failure with high mortality. Kallistatin, an endogenous tissue kallikrein inhibitor, has been reported to be protective in various human diseases. The aim of this study was to assess the correlations of kallistatin with other biomarkers and to determine whether kallistatin levels have a prognostic value in severe CAP.MethodsPlasma samples and clinical data were prospectively collected from 54 patients with severe CAP requiring ICU admission. Seventeen healthy control subjects were included for comparison. Plasma kallistatin, kallikrein, and other biomarkers of inflammation (tumor necrosis factor-α (TNF-α), interleukin (IL)-1β, IL-6, IL-8, C-reactive protein (CRP)), and anti-coagulation (protein C, anti-thrombin III) were measured on days 1 and 4 of ICU admission. Comparison between survivors (n = 41) and nonsurvivors (n = 13) was performed.ResultsPlasma kallistatin was significantly consumed in severe CAP patients compared with healthy individuals. Lower day 1 kallistatin levels showed a strong trend toward increased mortality (P = 0.018) and higher day 1 CURB-65 scores (P = 0.004). Plasma kallistatin levels on day 1 of ICU admission were significantly decreased in patients who developed septic shock (P = 0.017) and who had acute respiratory distress syndrome (P = 0.044). In addition, kallistatin levels were positively correlated with anti-thrombin III and protein C and inversely correlated with IL-1β, IL-6, and CRP levels. In a multivariate logistic regression analysis, higher day 1 CURB-65 scores were independent predictors of mortality (odds ratio = 29.9; P = 0.009). Also, higher day 1 kallistatin levels were independently associated with a decreased risk of death (odds ratio, 0.1) with a nearly significant statistical difference (P = 0.056). Furthermore, we found that a cutoff level of 6.5 μg/ml of day 1 kallistatin determined by receiver operating characteristic curves could be used to distinguish between patients who survived in 60 days and those who did not.ConclusionsThese results suggest that kallistatin may serve as a novel marker for severe CAP prognosis and may be involved in the pathogenesis of CAP through antiinflammatory and anticoagulation effects.

Dataset Information

Using data-driven rules to predict mortality in severe community acquired pneumonia.

Publications

Using data-driven rules to predict mortality in severe community acquired pneumonia.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure